In [1]:
import pandas as pd
import numpy as np
import scipy.stats as stats
import matplotlib.pyplot as plt
import seaborn as sns
import plotly.express as px
import plotly.io as pio
pio.renderers.default = 'notebook_connected'
In [2]:
# Load the datasets
customer_data = pd.read_csv('Customers.csv')
sales_data = pd.read_csv('Sales.csv')
In [3]:
# Display the customer data
customer_data
Out[3]:
Customer ID bracket_cust_id Customer Name Segment Age Country City State Postal Code Region
0 CG-12520 (CG-12520) Claire Gute Consumer 67 United States Henderson Kentucky 42420 South
1 DV-13045 (DV-13045) Darrin Van Huff Corporate 31 United States Los Angeles California 90036 West
2 SO-20335 (SO-20335) Sean O'Donnell Consumer 65 United States Fort Lauderdale Florida 33311 South
3 BH-11710 (BH-11710) Brosina Hoffman Consumer 20 United States Los Angeles California 90032 West
4 AA-10480 (AA-10480) Andrew Allen Consumer 50 United States Concord North Carolina 28027 South
... ... ... ... ... ... ... ... ... ... ...
788 CJ-11875 (CJ-11875) Carl Jackson Corporate 64 United States Philadelphia Pennsylvania 19140 East
789 RS-19870 (RS-19870) Roy Skaria Home Office 39 United States Burlington Iowa 52601 Central
790 SC-20845 (SC-20845) Sung Chung Consumer 69 United States Arlington Heights Illinois 60004 Central
791 RE-19405 (RE-19405) Ricardo Emerson Consumer 35 United States Kent Ohio 44240 East
792 SM-20905 (SM-20905) Susan MacKendrick Consumer 46 United States Newark Ohio 43055 East

793 rows × 10 columns

In [4]:
# Display the sales data
sales_data
Out[4]:
Order Line Order ID Order Date Ship Date Ship Mode Customer ID Product ID Sales Quantity Discount Profit
0 1 CA-2016-152156 11/8/2016 11/11/2016 Second Class CG-12520 FUR-BO-10001798 261.96 2 0.00 41.91
1 2 CA-2016-152156 11/8/2016 11/11/2016 Second Class CG-12520 FUR-CH-10000454 731.94 3 0.00 219.58
2 3 CA-2016-138688 6/12/2016 6/16/2016 Second Class DV-13045 OFF-LA-10000240 14.62 2 0.00 6.87
3 4 US-2015-108966 10/11/2015 10/18/2015 Standard Class SO-20335 FUR-TA-10000577 957.58 5 0.45 -383.03
4 5 US-2015-108966 10/11/2015 10/18/2015 Standard Class SO-20335 OFF-ST-10000760 22.37 2 0.20 2.52
... ... ... ... ... ... ... ... ... ... ... ...
9989 9990 CA-2014-110422 1/21/2014 1/23/2014 Second Class TB-21400 FUR-FU-10001889 25.25 3 0.20 4.10
9990 9991 CA-2017-121258 2/26/2017 3/3/2017 Standard Class DB-13060 FUR-FU-10000747 91.96 2 0.00 15.63
9991 9992 CA-2017-121258 2/26/2017 3/3/2017 Standard Class DB-13060 TEC-PH-10003645 258.58 2 0.20 19.39
9992 9993 CA-2017-121258 2/26/2017 3/3/2017 Standard Class DB-13060 OFF-PA-10004041 29.60 4 0.00 13.32
9993 9994 CA-2017-119914 5/4/2017 5/9/2017 Second Class CC-12220 OFF-AP-10002684 243.16 2 0.00 72.95

9994 rows × 11 columns

Understanding the Data¶

In [5]:
# Check for missing values in customer data
customer_data.isnull().sum()
Out[5]:
Customer ID        0
bracket_cust_id    0
Customer Name      0
Segment            0
Age                0
Country            3
City               2
State              1
Postal Code        0
Region             0
dtype: int64
In [6]:
# Check for missing values in sales data
sales_data.isnull().sum()
Out[6]:
Order Line     0
Order ID       0
Order Date     0
Ship Date      0
Ship Mode      0
Customer ID    0
Product ID     0
Sales          0
Quantity       0
Discount       0
Profit         0
dtype: int64
In [7]:
# Drop duplicates if any
customer_data = customer_data.drop_duplicates()
sales_data = sales_data.drop_duplicates()
In [8]:
# Check data information for customer data
customer_data.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 793 entries, 0 to 792
Data columns (total 10 columns):
 #   Column           Non-Null Count  Dtype 
---  ------           --------------  ----- 
 0   Customer ID      793 non-null    object
 1   bracket_cust_id  793 non-null    object
 2   Customer Name    793 non-null    object
 3   Segment          793 non-null    object
 4   Age              793 non-null    int64 
 5   Country          790 non-null    object
 6   City             791 non-null    object
 7   State            792 non-null    object
 8   Postal Code      793 non-null    int64 
 9   Region           793 non-null    object
dtypes: int64(2), object(8)
memory usage: 62.1+ KB
In [9]:
# Check information for sales data
sales_data.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 9994 entries, 0 to 9993
Data columns (total 11 columns):
 #   Column       Non-Null Count  Dtype  
---  ------       --------------  -----  
 0   Order Line   9994 non-null   int64  
 1   Order ID     9994 non-null   object 
 2   Order Date   9994 non-null   object 
 3   Ship Date    9994 non-null   object 
 4   Ship Mode    9994 non-null   object 
 5   Customer ID  9994 non-null   object 
 6   Product ID   9994 non-null   object 
 7   Sales        9994 non-null   float64
 8   Quantity     9994 non-null   int64  
 9   Discount     9994 non-null   float64
 10  Profit       9994 non-null   float64
dtypes: float64(3), int64(2), object(6)
memory usage: 859.0+ KB

Cleaning and Preparing the Data¶

In [10]:
# Create categories based on Product ID (first 3 characters for category and next 2 for sub-category)
# Note: 'Product ID' exists in sales_data, not in customer_data — create the columns on sales_data.
sales_data['Category'] = sales_data['Product ID'].str[:3]
sales_data['Sub-Category'] = sales_data['Product ID'].str[4:6]
sales_data.head()
Out[10]:
Order Line Order ID Order Date Ship Date Ship Mode Customer ID Product ID Sales Quantity Discount Profit Category Sub-Category
0 1 CA-2016-152156 11/8/2016 11/11/2016 Second Class CG-12520 FUR-BO-10001798 261.96 2 0.00 41.91 FUR BO
1 2 CA-2016-152156 11/8/2016 11/11/2016 Second Class CG-12520 FUR-CH-10000454 731.94 3 0.00 219.58 FUR CH
2 3 CA-2016-138688 6/12/2016 6/16/2016 Second Class DV-13045 OFF-LA-10000240 14.62 2 0.00 6.87 OFF LA
3 4 US-2015-108966 10/11/2015 10/18/2015 Standard Class SO-20335 FUR-TA-10000577 957.58 5 0.45 -383.03 FUR TA
4 5 US-2015-108966 10/11/2015 10/18/2015 Standard Class SO-20335 OFF-ST-10000760 22.37 2 0.20 2.52 OFF ST
In [11]:
# Replace missing cities/states based on postal codes with the most frequent city/state for that postal code
# Create mapping: postal_code -> {'State': state_value, 'City': city_value}
postal_code_to_city = customer_data.groupby('Postal Code')[['Country', 'State', 'City']].agg(lambda x: x.mode().iloc[0] if not x.mode().empty else np.nan).to_dict(orient='index')

# Fill missing City/State using the mapping
def _fill_city_state(row):
    info = postal_code_to_city.get(row['Postal Code'], {})
    city = row['City'] if pd.notnull(row['City']) else info.get('City', np.nan)
    state = row['State'] if pd.notnull(row['State']) else info.get('State', np.nan)
    country = row['Country'] if pd.notnull(row['Country']) else info.get('Country', np.nan)
    return pd.Series({'City': city, 'State': state, 'Country': country})

filled = customer_data.apply(_fill_city_state, axis=1)
customer_data['City'] = filled['City']
customer_data['State'] = filled['State']
customer_data['Country'] = filled['Country']

customer_data.isnull().sum()
Out[11]:
Customer ID        0
bracket_cust_id    0
Customer Name      0
Segment            0
Age                0
Country            1
City               0
State              0
Postal Code        0
Region             0
dtype: int64
In [12]:
# Descriptive statistics for Age
customer_data['Age'].describe()
Out[12]:
count    793.000000
mean      44.467844
std       15.526673
min       18.000000
25%       31.000000
50%       45.000000
75%       58.000000
max       70.000000
Name: Age, dtype: float64
In [13]:
# Age categorization from 18 to 71 in 4 bins
age_bins = [18, 30, 45, 60, 71]
age_labels = ['18-30', '31-45', '46-60', '61+']
customer_data['Age Group'] = pd.cut(customer_data['Age'], bins=age_bins, labels=age_labels, right=False)
customer_data['Age Group'].value_counts().sort_index()
Out[13]:
Age Group
18-30    165
31-45    228
46-60    219
61+      181
Name: count, dtype: int64
In [14]:
# Drop country column because it has only one value
customer_data = customer_data.drop(columns=['Country'])
In [15]:
# The range of dates in the sales data
sales_data['Order Date'] = pd.to_datetime(sales_data['Order Date']).dt.date
print(f"Sales data ranges from {sales_data['Order Date'].min()} to {sales_data['Order Date'].max()}, totaling {(sales_data['Order Date'].max() - sales_data['Order Date'].min()).days} days.")

Sales data ranges from 2014-01-03 to 2017-12-30, totaling 1457 days.
In [16]:
# Replace category codes with descriptive names
category_mapping = {'OFF': 'Office Supplies', 'TEC': 'Technology', 'FUR': 'Furniture'}
sales_data['Category'] = sales_data['Category'].map(category_mapping)
sales_data.head()
Out[16]:
Order Line Order ID Order Date Ship Date Ship Mode Customer ID Product ID Sales Quantity Discount Profit Category Sub-Category
0 1 CA-2016-152156 2016-11-08 11/11/2016 Second Class CG-12520 FUR-BO-10001798 261.96 2 0.00 41.91 Furniture BO
1 2 CA-2016-152156 2016-11-08 11/11/2016 Second Class CG-12520 FUR-CH-10000454 731.94 3 0.00 219.58 Furniture CH
2 3 CA-2016-138688 2016-06-12 6/16/2016 Second Class DV-13045 OFF-LA-10000240 14.62 2 0.00 6.87 Office Supplies LA
3 4 US-2015-108966 2015-10-11 10/18/2015 Standard Class SO-20335 FUR-TA-10000577 957.58 5 0.45 -383.03 Furniture TA
4 5 US-2015-108966 2015-10-11 10/18/2015 Standard Class SO-20335 OFF-ST-10000760 22.37 2 0.20 2.52 Office Supplies ST
In [17]:
# Replace sub-category codes with descriptive names ['BO', 'CH', 'LA', 'TA', 'ST', 'FU', 'AR', 'PH', 'BI', 'AP', 'PA','AC', 'EN', 'FA', 'SU', 'MA', 'CO']
sub_category_mapping = {
    'BO': 'Books',
    'CH': 'Chairs',
    'LA': 'Labels',
    'TA': 'Tables',
    'ST': 'Storage',
    'FU': 'Furnishings',
    'AR': 'Art',
    'PH': 'Phones',
    'BI': 'Binders',
    'AP': 'Appliances',
    'PA': 'Paper',
    'AC': 'Accessories',
    'EN': 'Envelopes',
    'FA': 'Fasteners',
    'SU': 'Supplies',
    'MA': 'Machines',
    'CO': 'Copiers'
}
sales_data['Sub-Category'] = sales_data['Sub-Category'].map(sub_category_mapping)
sales_data.head()
Out[17]:
Order Line Order ID Order Date Ship Date Ship Mode Customer ID Product ID Sales Quantity Discount Profit Category Sub-Category
0 1 CA-2016-152156 2016-11-08 11/11/2016 Second Class CG-12520 FUR-BO-10001798 261.96 2 0.00 41.91 Furniture Books
1 2 CA-2016-152156 2016-11-08 11/11/2016 Second Class CG-12520 FUR-CH-10000454 731.94 3 0.00 219.58 Furniture Chairs
2 3 CA-2016-138688 2016-06-12 6/16/2016 Second Class DV-13045 OFF-LA-10000240 14.62 2 0.00 6.87 Office Supplies Labels
3 4 US-2015-108966 2015-10-11 10/18/2015 Standard Class SO-20335 FUR-TA-10000577 957.58 5 0.45 -383.03 Furniture Tables
4 5 US-2015-108966 2015-10-11 10/18/2015 Standard Class SO-20335 OFF-ST-10000760 22.37 2 0.20 2.52 Office Supplies Storage
In [18]:
# Extract month and year from 'Order Date'
sales_data['Month'] = pd.to_datetime(sales_data['Order Date']).dt.month
sales_data['Year'] = pd.to_datetime(sales_data['Order Date']).dt.year
In [19]:
# Number of orders in sales data
print(f"Total number of orders: {sales_data['Order ID'].nunique()}")

Total number of orders: 5009
In [20]:
# Number of customers ordering
print(f"Total number of customers: {sales_data['Customer ID'].nunique()}")

Total number of customers: 793
In [21]:
# Number of unique products in sales data
print(f"Number of unique products: {sales_data['Product ID'].nunique()}")

Number of unique products: 1862
In [22]:
# Number of unique categories
print(f"Number of unique categories: {sales_data['Category'].nunique()}")
print(f"Number of unique sub-categories: {sales_data['Sub-Category'].nunique()}")

Number of unique categories: 3
Number of unique sub-categories: 17
In [23]:
# Count of categories in sales data
print(sales_data['Category'].value_counts(normalize=True))# Count of sub-categories in sales data
print(sales_data['Sub-Category'].value_counts(normalize=True))

Category
Office Supplies    0.602962
Furniture          0.212227
Technology         0.184811
Name: proportion, dtype: float64
Sub-Category
Binders        0.152391
Paper          0.137082
Furnishings    0.095757
Phones         0.088953
Storage        0.084651
Art            0.079648
Accessories    0.077547
Chairs         0.061737
Appliances     0.046628
Labels         0.036422
Tables         0.031919
Envelopes      0.025415
Books          0.022814
Fasteners      0.021713
Supplies       0.019011
Machines       0.011507
Copiers        0.006804
Name: proportion, dtype: float64
In [24]:
# Profit or loss
sales_data['Profit/Loss'] = np.where(sales_data['Profit'] >= 0, 'Profit', 'Loss')
print(sales_data['Profit/Loss'].value_counts(normalize=True))
# Visalize profit vs loss
plt.figure(figsize=(8, 5))
sns.countplot(data=sales_data, x='Profit/Loss', palette='Set2', hue='Profit/Loss')
plt.title('Profit vs Loss')
plt.xlabel('Profit/Loss')
plt.ylabel('Count')
plt.show()

Profit/Loss
Profit    0.812788
Loss      0.187212
Name: proportion, dtype: float64
No description has been provided for this image

Sales data ranges from 2014-01-03 to 2017-12-30, totaling 1457 days. There are 793 unique customers, 5009 unique orders, and 1862 unique products in the sales dataset. The customer data includes 793 unique customers across various segments and regions. The sales data contains 3 main categories: Office Supplies, Technology, and Furniture, with 17 sub-categories under each. Profitability analysis shows that approximately 81.28% of transactions are profitable, while 18.72% result in a loss.

In [25]:
# Number of unique customers in customer data
print(f"Number of unique customers: {customer_data['Customer ID'].nunique()}")

Number of unique customers: 793
In [26]:
# Customer segments in customer data
print(customer_data['Segment'].value_counts(normalize=True))
# Visualize customer segments
plt.figure(figsize=(8, 5))
sns.countplot(data=customer_data, x='Segment', palette='Set2', hue='Segment')
plt.title('Customer Distribution by Segment')
plt.xlabel('Segment')
plt.ylabel('Count')
plt.show()

Segment
Consumer       0.515763
Corporate      0.297604
Home Office    0.186633
Name: proportion, dtype: float64
No description has been provided for this image
In [27]:
# Regions in customer data
print("Customer counts by region:")
print(customer_data['Region'].value_counts())
print("\nProportion of customers by region:")
print(customer_data['Region'].value_counts(normalize=True))
# Visualize regions
plt.figure(figsize=(8, 5))
sns.countplot(data=customer_data, x='Region', palette='Set2', hue='Region')
plt.title('Customer Distribution by Region')
plt.xlabel('Region')
plt.ylabel('Count')
plt.show()

Customer counts by region:
Region
West       255
East       220
Central    184
South      134
Name: count, dtype: int64

Proportion of customers by region:
Region
West       0.321564
East       0.277427
Central    0.232030
South      0.168979
Name: proportion, dtype: float64
No description has been provided for this image
  • In the customer data, there are a total of 793 unique customers in three segments: Consumer, Corporate, and Home Office. The Consumer segment makes up the largest proportion of customers at approximately 51.58%, followed by Corporate at 29.76%, and Home Office at 18.66%.
  • The regions are distributed as follows: West (32.16%), East (27.74%), Central (23.20%), and South (16.90%).
In [28]:
# Merge customer and sales data on 'Customer ID'
merged_data = pd.merge(sales_data, customer_data, on='Customer ID', how='left')
merged_data.head()
Out[28]:
Order Line Order ID Order Date Ship Date Ship Mode Customer ID Product ID Sales Quantity Discount ... Profit/Loss bracket_cust_id Customer Name Segment Age City State Postal Code Region Age Group
0 1 CA-2016-152156 2016-11-08 11/11/2016 Second Class CG-12520 FUR-BO-10001798 261.96 2 0.00 ... Profit (CG-12520) Claire Gute Consumer 67 Henderson Kentucky 42420 South 61+
1 2 CA-2016-152156 2016-11-08 11/11/2016 Second Class CG-12520 FUR-CH-10000454 731.94 3 0.00 ... Profit (CG-12520) Claire Gute Consumer 67 Henderson Kentucky 42420 South 61+
2 3 CA-2016-138688 2016-06-12 6/16/2016 Second Class DV-13045 OFF-LA-10000240 14.62 2 0.00 ... Profit (DV-13045) Darrin Van Huff Corporate 31 Los Angeles California 90036 West 31-45
3 4 US-2015-108966 2015-10-11 10/18/2015 Standard Class SO-20335 FUR-TA-10000577 957.58 5 0.45 ... Loss (SO-20335) Sean O'Donnell Consumer 65 Fort Lauderdale Florida 33311 South 61+
4 5 US-2015-108966 2015-10-11 10/18/2015 Standard Class SO-20335 OFF-ST-10000760 22.37 2 0.20 ... Profit (SO-20335) Sean O'Donnell Consumer 65 Fort Lauderdale Florida 33311 South 61+

5 rows × 25 columns

Exploring Key Business Questions¶

Customer Insights¶

Age Distribution Analysis¶

In [29]:
# Age distribution
plt.figure(figsize=(8, 5))
sns.histplot(customer_data['Age'], bins=30, kde=True)
plt.title('Age Distribution of Customers')
plt.xlabel('Age')
plt.ylabel('Frequency')
plt.show()
No description has been provided for this image
In [30]:
# Box plot of age
age_box_fig = px.box(customer_data, y='Age',height=600, width=800, title='Box Plot of Customer Ages')
age_box_fig.show()
In [31]:
# What is the age distribution across different customer segments?
age_segment_fig = px.box(customer_data, x='Segment', y='Age', color='Segment',height=600, width=800, title='Age Distribution Across Customer Segments')
age_segment_fig.show()
In [32]:
# ANOVA test to see if age differs significantly across customer segments

anova_result = stats.f_oneway(
    customer_data[customer_data['Segment'] == 'Consumer']['Age'].dropna(),
    customer_data[customer_data['Segment'] == 'Corporate']['Age'].dropna(),
    customer_data[customer_data['Segment'] == 'Home Office']['Age'].dropna()
)
print(f"ANOVA result: F-statistic = {anova_result.statistic}, p-value = {anova_result.pvalue}")
if anova_result.pvalue < 0.05:
    print("There is a significant difference in age across customer segments.")
else:
    print("No significant difference in age across customer segments.")

ANOVA result: F-statistic = 0.1795113122651887, p-value = 0.8357125753930718
No significant difference in age across customer segments.

Box plot shows similar age distributions across all customer segments, with median ages around the same level and overlapping interquartile ranges. The ANOVA test results (p-value > 0.05) indicate that there is no statistically significant difference in age across customer segments in this dataset.

In [33]:
# What is the age distribution across different customer regions?
age_region_fig = px.box(customer_data, x='Region', y='Age', color='Region',height=600, width=800, title='Age Distribution Across Customer Regions')
age_region_fig.show()
In [34]:
# ANOVA test to see if age differs significantly across customer regions
anova_region_result = stats.f_oneway(
    customer_data[customer_data['Region'] == 'East']['Age'].dropna(),
    customer_data[customer_data['Region'] == 'West']['Age'].dropna(),
    customer_data[customer_data['Region'] == 'Central']['Age'].dropna()
)
print(f"ANOVA result: F-statistic = {anova_region_result.statistic}, p-value = {anova_region_result.pvalue}")
if anova_region_result.pvalue < 0.05:
    print("There is a significant difference in age across customer regions.")
else:
    print("No significant difference in age across customer regions.")

ANOVA result: F-statistic = 0.02358030089268894, p-value = 0.9766963698111525
No significant difference in age across customer regions.
In [35]:
# What is the age distribution across different State categories?
age_region_fig = px.box(customer_data, x='State', y='Age', color='Region',height=800, width=1000, title='Age Distribution Across Customer Regions')
age_region_fig.show()
In [36]:
# ANOVA test to see if age differs significantly across State categories
state_groups = [customer_data[customer_data['State'] == state]['Age'].dropna() for state in customer_data['State'].unique()]
anova_state_result = stats.f_oneway(*state_groups)
print(f"ANOVA result: F-statistic = {anova_state_result.statistic}, p-value = {anova_state_result.pvalue}")
if anova_state_result.pvalue < 0.05:
    print("There is a significant difference in age across State categories.")
else:
    print("No significant difference in age across State categories.")

ANOVA result: F-statistic = 1.0892692031440407, p-value = 0.3279364695492146
No significant difference in age across State categories.
In [37]:
# What is the age distribution across different Ship Modes?
age_shipmode_fig = px.box(merged_data, x='Ship Mode', y='Age', color='Ship Mode',height=600, width=800, title='Age Distribution Across Ship Modes')
age_shipmode_fig.show()
In [38]:
# ANOVA test to see if age differs significantly across Ship Modes
anova_shipmode_result = stats.f_oneway(
    merged_data[merged_data['Ship Mode'] == 'First Class']['Age'].dropna(),
    merged_data[merged_data['Ship Mode'] == 'Second Class']['Age'].dropna(),
    merged_data[merged_data['Ship Mode'] == 'Standard Class']['Age'].dropna(),
    merged_data[merged_data['Ship Mode'] == 'Same Day']['Age'].dropna()
)
print(f"ANOVA result: F-statistic = {anova_shipmode_result.statistic}, p-value = {anova_shipmode_result.pvalue}")
if anova_shipmode_result.pvalue < 0.05:
    print("There is a significant difference in age across Ship Modes.")
else:
    print("No significant difference in age across Ship Modes.")

ANOVA result: F-statistic = 2.3584917500937133, p-value = 0.06959702336876959
No significant difference in age across Ship Modes.
In [39]:
# What is the age distribution across different Categories?
age_category_fig = px.box(merged_data, x='Category', y='Age', color='Category',height=600, width=800, title='Age Distribution Across Categories')
age_category_fig.show()
In [40]:
# ANOVA test to see if age differs significantly across categories
anova_category_result = stats.f_oneway(
    merged_data[merged_data['Category'] == 'Furniture']['Age'].dropna(),
    merged_data[merged_data['Category'] == 'Office Supplies']['Age'].dropna(),
    merged_data[merged_data['Category'] == 'Technology']['Age'].dropna()
)
print(f"ANOVA result: F-statistic = {anova_category_result.statistic}, p-value = {anova_category_result.pvalue}")
if anova_category_result.pvalue < 0.05:
    print("There is a significant difference in age across Categories.")
else:
    print("No significant difference in age across Categories.")

ANOVA result: F-statistic = 0.6428165301824491, p-value = 0.5258311238049834
No significant difference in age across Categories.
In [41]:
# What is the age distribution across different Sub Categories?
age_subcategory_fig = px.box(merged_data, x='Sub-Category', y='Age', color='Sub-Category', height=600, width=800, title='Age Distribution Across Sub Categories')
age_subcategory_fig.show()
In [42]:
# ANOVA test to see if age differs significantly across sub categories
age_value_groups = [merged_data[merged_data['Sub-Category'] == subcategory]['Age'] for subcategory in merged_data['Sub-Category'].unique()]
anova_age_value_result = stats.f_oneway(*age_value_groups)
print(f"ANOVA result: F-statistic = {anova_age_value_result.statistic}, p-value = {anova_age_value_result.pvalue}")
if anova_age_value_result.pvalue < 0.05:
    print("There is a significant difference in age across Sub Categories.")
else:
    print("No significant difference in age across Sub Categories.")

ANOVA result: F-statistic = 0.7630266780316128, p-value = 0.7294351044889795
No significant difference in age across Sub Categories.
In [43]:
# What is the age distribution across different Profit/Loss categories?
age_subcategory_fig = px.box(merged_data, x='Profit/Loss', y='Age', color='Profit/Loss', height=600, width=800, title='Age Distribution Across Profit/Loss Categories')
age_subcategory_fig.show()
In [44]:
# T-test test to see if age differs significantly across Profit/Loss categories
profit_ages = merged_data[merged_data['Profit/Loss'] == 'Profit']['Age'].dropna()
loss_ages = merged_data[merged_data['Profit/Loss'] == 'Loss']['Age'].dropna()
t_test_result = stats.ttest_ind(profit_ages, loss_ages, equal_var=False)
print(f"T-test result: t-statistic = {t_test_result.statistic}, p-value = {t_test_result.pvalue}")
if t_test_result.pvalue < 0.05:
    print("There is a significant difference in age between Profit and Loss categories.")
else:
    print("No significant difference in age between Profit and Loss categories.")

T-test result: t-statistic = 1.7073428090393807, p-value = 0.08786954739035212
No significant difference in age between Profit and Loss categories.

There is not significant difference in age across customer ship modes, segments, regions, categories, sub-categories, and profit/loss categories. In all cases, the p-values from ANOVA or T-tests are greater than 0.05, indicating that age distributions are similar across these different groupings. This suggests that age is not a distinguishing factor for customers based on these attributes in this dataset.

Sales Analysis¶

In [45]:
# What is the sales distribution across different customer segments?
sales_segment_fig = px.box(merged_data, x='Segment', y='Sales', color='Segment', height=600, width=800, title='Sales Distribution Across Customer Segments')
sales_segment_fig.show()
# Descriptive statistics for sales cross customer segments
merged_data.groupby('Segment')['Sales'].describe()
Out[45]:
count mean std min 25% 50% 75% max
Segment
Consumer 5191.0 223.733643 585.516211 0.44 17.2650 53.72 206.105 13999.96
Corporate 3020.0 233.823325 599.410418 0.56 17.4675 56.54 212.895 17499.95
Home Office 1783.0 240.972120 755.516423 0.99 17.0450 52.44 210.905 22638.48
In [46]:
# ANOVA test to see if sales differ significantly across customer segments
anova_sales_segment_result = stats.f_oneway(
    merged_data[merged_data['Segment'] == 'Consumer']['Sales'].dropna(),
    merged_data[merged_data['Segment'] == 'Corporate']['Sales'].dropna(),
    merged_data[merged_data['Segment'] == 'Home Office']['Sales'].dropna()
)
print(f"ANOVA result: F-statistic = {anova_sales_segment_result.statistic}, p-value = {anova_sales_segment_result.pvalue}")
if anova_sales_segment_result.pvalue < 0.05:
    print("There is a significant difference in sales across customer segments.")
else:
    print("No significant difference in sales across customer segments.")

ANOVA result: F-statistic = 0.5952018176125338, p-value = 0.5514708144500259
No significant difference in sales across customer segments.
In [47]:
# What is the sales distribution across different customer regions?
sales_region_fig = px.box(merged_data, x='Region', y='Sales', color='Region', height=600, width=800, title='Sales Distribution Across Customer Regions')
sales_region_fig.show()
# Descriptive statistics for sales cross customer regions
merged_data.groupby('Region')['Sales'].describe()
Out[47]:
count mean std min 25% 50% 75% max
Region
Central 2371.0 218.810654 536.744807 0.84 16.1900 54.480 207.3800 10499.97
East 2762.0 221.482448 614.889173 0.56 16.9225 48.900 191.8225 13999.96
South 1626.0 247.252189 768.940813 0.98 17.4625 54.565 239.8875 22638.48
West 3235.0 236.363039 607.368739 0.44 17.9400 59.980 215.0500 17499.95
In [48]:
# What is the sales distribution across different State categories?
sales_region_fig = px.box(merged_data, x='State', y='Sales', color='Region', height=800, width=1000, color_discrete_sequence=px.colors.qualitative.Set1, title='Sales Distribution Across Customer Regions')
sales_region_fig.show()
# Descriptive statistics for sales cross State categories
merged_data.groupby('State')['Sales'].describe()
Out[48]:
count mean std min 25% 50% 75% max
State
Alabama 127.0 244.401654 426.151092 1.58 16.6650 49.560 323.9600 3040.00
Arizona 301.0 272.379103 767.047298 2.03 17.9400 69.280 243.9900 9892.74
Arkansas 21.0 218.217619 369.517428 4.37 8.9500 47.990 105.4200 1080.10
California 1998.0 225.744054 485.623164 0.85 17.9400 57.580 216.2925 9449.95
Colorado 283.0 203.261767 358.003577 1.37 18.5600 64.850 229.9450 2563.06
Connecticut 74.0 250.672973 564.645921 1.64 14.5550 47.935 231.8475 4164.05
Delaware 93.0 226.409140 431.949342 1.48 20.1000 68.110 226.5600 2888.13
District of Columbia 11.0 199.860000 402.031815 8.23 16.0700 26.340 77.8350 1287.45
Florida 290.0 172.424310 292.195120 1.17 15.5500 45.680 180.0375 2177.58
Georgia 201.0 255.722090 624.428359 1.78 17.4700 59.760 232.8800 6354.95
Illinois 574.0 196.550052 483.438906 1.19 14.9475 50.675 215.2650 8187.65
Indiana 167.0 137.592096 193.229214 1.87 15.5400 45.680 176.5400 931.18
Iowa 45.0 276.137111 513.463035 5.40 21.2400 68.520 287.8800 2548.56
Kansas 13.0 132.896154 222.719650 7.22 25.9000 54.710 122.3800 848.54
Kentucky 149.0 195.596309 326.734453 2.90 22.6800 69.380 199.9800 2003.17
Louisiana 85.0 195.598353 413.837515 2.41 13.9600 42.210 152.6900 2879.95
Maryland 24.0 66.200833 78.450988 1.81 14.4000 36.510 91.8575 355.96
Massachusetts 114.0 249.222368 595.462590 1.08 23.9975 56.625 241.5600 5399.91
Michigan 220.0 263.985636 563.072625 1.34 21.3525 81.775 241.4550 4297.64
Minnesota 167.0 299.774731 615.538807 2.29 18.3900 65.520 241.0950 4404.90
Mississippi 61.0 317.168033 460.995262 2.07 28.7900 101.990 421.1000 2025.36
Missouri 60.0 216.886000 350.859611 3.11 16.9475 39.120 290.7550 1577.94
Nebraska 32.0 202.887500 459.901810 1.98 20.9250 69.955 167.9675 2479.96
Nevada 12.0 101.249167 156.906351 8.29 18.8575 52.755 77.7325 546.06
New Hampshire 28.0 347.147857 653.815815 5.21 25.4250 73.490 237.1975 2973.32
New Jersey 123.0 164.773171 319.968129 2.37 18.5700 46.260 131.4350 2239.94
New Mexico 41.0 147.468049 222.849299 3.10 20.2400 47.500 127.9500 862.34
New York 1110.0 251.846775 823.242597 0.56 17.2500 51.070 189.4575 13999.96
North Carolina 343.0 340.045160 1435.883562 1.25 15.6050 51.550 241.3050 22638.48
Ohio 480.0 155.773750 291.857503 1.39 16.1500 38.960 154.5900 2807.84
Oklahoma 16.0 320.988750 556.661844 11.76 34.8350 66.370 379.1250 2154.90
Oregon 60.0 288.786167 505.332267 0.44 15.4825 75.210 234.5300 2799.94
Pennsylvania 641.0 222.837114 443.426351 1.04 15.7600 56.060 220.7000 4899.93
Rhode Island 64.0 199.732656 482.399746 1.80 14.6500 40.815 152.0750 3357.60
South Dakota 16.0 146.225000 164.203397 5.95 15.0850 61.505 205.9050 492.77
Tennessee 167.0 219.961677 382.732547 5.02 25.0700 69.220 209.6450 2453.43
Texas 917.0 210.205256 594.259633 0.84 15.2800 46.860 183.3700 10499.97
Utah 80.0 204.038000 428.041798 1.91 19.1275 47.040 156.0500 2321.90
Virginia 182.0 255.611209 548.723571 0.98 16.8150 52.595 262.2450 4899.93
Washington 460.0 289.515717 1016.844663 0.99 17.4525 68.140 207.9975 17499.95
Wisconsin 144.0 284.515347 601.098892 2.63 23.5475 80.195 270.5250 4305.55
In [49]:
# ANOVA test to see if sales differs significantly across customer regions
anova_sales_region_result = stats.f_oneway(
    merged_data[merged_data['Region'] == 'East']['Sales'].dropna(),
    merged_data[merged_data['Region'] == 'West']['Sales'].dropna(),
    merged_data[merged_data['Region'] == 'Central']['Sales'].dropna()
)
print(f"ANOVA result: F-statistic = {anova_sales_region_result.statistic}, p-value = {anova_sales_region_result.pvalue}")
if anova_sales_region_result.pvalue < 0.05:
    print("There is a significant difference in sales across customer regions.")
else:
    print("No significant difference in sales across customer regions.")

ANOVA result: F-statistic = 0.751235202598438, p-value = 0.47181527122479683
No significant difference in sales across customer regions.
In [50]:
# What is the sales distribution across different Ship Modes?
sales_ship_mode_fig = px.box(merged_data, x='Ship Mode', y='Sales', color='Ship Mode', height=600, width=800, title='Sales Distribution Across Ship Modes')
sales_ship_mode_fig.show()
# Descriptive statistics for sales cross ship modes
merged_data.groupby('Ship Mode')['Sales'].describe()
Out[50]:
count mean std min 25% 50% 75% max
Ship Mode
First Class 1538.0 228.497029 630.384744 0.98 18.165 54.94 199.9800 13999.96
Same Day 543.0 236.396169 555.258823 0.56 16.615 55.18 209.7950 7999.98
Second Class 1945.0 236.089172 559.146855 1.19 17.880 57.90 223.9200 8749.95
Standard Class 5968.0 227.583123 646.739021 0.44 16.780 52.77 209.9475 22638.48
In [51]:
# ANOVA test to see if sales differs significantly across Ship Modes
anova_sales_ship_mode_result = stats.f_oneway(
    merged_data[merged_data['Ship Mode'] == 'First Class']['Sales'].dropna(),
    merged_data[merged_data['Ship Mode'] == 'Second Class']['Sales'].dropna(),
    merged_data[merged_data['Ship Mode'] == 'Standard Class']['Sales'].dropna()
)
print(f"ANOVA result: F-statistic = {anova_sales_ship_mode_result.statistic}, p-value = {anova_sales_ship_mode_result.pvalue}")
if anova_sales_ship_mode_result.pvalue < 0.05:
    print("There is a significant difference in sales across Ship Modes.")
else:
    print("No significant difference in sales across Ship Modes.")

ANOVA result: F-statistic = 0.13725330616108347, p-value = 0.8717511168326189
No significant difference in sales across Ship Modes.
In [52]:
# What is the sales distribution across different Categories?
sales_category_fig = px.box(merged_data, x='Category', y='Sales', color='Category', height=600, width=800, title='Sales Distribution Across Categories')
sales_category_fig.show()
# Descriptive statistics for sales cross categories
merged_data.groupby('Category')['Sales'].describe()
Out[52]:
count mean std min 25% 50% 75% max
Category
Furniture 2121.0 349.834974 503.179147 1.89 47.040 182.22 435.170 4416.17
Office Supplies 6026.0 119.324094 382.182252 0.44 11.760 27.42 79.920 9892.74
Technology 1847.0 452.709312 1108.655892 0.99 68.015 166.16 448.535 22638.48
In [53]:
# ANOVA test to see if sales differs significantly across categories
anova_sales_category_result = stats.f_oneway(
    merged_data[merged_data['Category'] == 'Furniture']['Sales'].dropna(),
    merged_data[merged_data['Category'] == 'Office Supplies']['Sales'].dropna(),
    merged_data[merged_data['Category'] == 'Technology']['Sales'].dropna()
)
print(f"ANOVA result: F-statistic = {anova_sales_category_result.statistic}, p-value = {anova_sales_category_result.pvalue}")
if anova_sales_category_result.pvalue < 0.05:
    print("There is a significant difference in sales across Categories.")
else:
    print("No significant difference in sales across Categories.")

ANOVA result: F-statistic = 265.4898596241737, p-value = 4.5571281511140176e-113
There is a significant difference in sales across Categories.
In [54]:
# What is the sales distribution across different Sub Categories?
sales_sub_category_fig = px.box(merged_data, x='Sub-Category', y='Sales', color='Sub-Category', height=600, width=800, title='Sales Distribution Across Sub Categories')
sales_sub_category_fig.show()
# Descriptive statistics for sales cross sub categories
merged_data.groupby('Sub-Category')['Sales'].describe()
Out[54]:
count mean std min 25% 50% 75% max
Sub-Category
Accessories 775.0 215.974594 334.965025 0.99 47.9800 100.000 239.9600 3347.37
Appliances 466.0 230.755665 388.949730 0.44 35.0550 82.690 241.4400 2625.12
Art 796.0 34.068844 60.122373 1.34 7.9800 15.505 36.4400 1113.02
Binders 1523.0 133.560584 563.251213 0.56 8.1150 18.560 51.8500 9892.74
Books 228.0 503.859868 638.748462 35.49 190.4975 306.820 522.2850 4404.90
Chairs 617.0 532.332464 550.148241 26.64 191.9600 362.140 662.8800 4416.17
Copiers 68.0 2198.941324 3175.666232 299.99 599.9875 1099.980 2399.9600 17499.95
Envelopes 254.0 64.867638 84.351785 1.63 15.2575 29.045 71.9475 604.66
Fasteners 217.0 13.936636 12.416608 1.24 5.6800 10.580 17.9000 93.36
Furnishings 957.0 95.825622 147.893645 1.89 19.3000 41.960 106.6800 1336.44
Labels 364.0 34.303022 74.119166 2.09 8.2600 14.940 29.2400 786.48
Machines 115.0 1645.553739 2765.101902 11.56 287.9400 599.990 2120.9400 22638.48
Paper 1370.0 57.284117 78.167675 3.38 14.5150 26.720 61.9600 733.95
Phones 889.0 371.211586 491.457448 2.97 84.7800 209.970 475.9400 4548.81
Storage 846.0 264.590532 355.222519 4.46 46.5300 113.930 340.2000 2934.33
Supplies 190.0 245.650105 923.828866 1.74 12.1250 27.930 55.4925 8187.65
Tables 319.0 648.795235 615.774358 24.37 244.0100 447.840 872.1700 4297.64
In [55]:
# ANOVA test to see if sales differs significantly across sub categories
sales_value_groups = [merged_data[merged_data['Sub-Category'] == subcategory]['Sales'] for subcategory in merged_data['Sub-Category'].unique()]
anova_sales_value_result = stats.f_oneway(*sales_value_groups)
print(f"ANOVA result: F-statistic = {anova_sales_value_result.statistic}, p-value = {anova_sales_value_result.pvalue}")
if anova_sales_value_result.pvalue < 0.05:
    print("There is a significant difference in sales across Sub Categories.")
else:
    print("No significant difference in sales across Sub Categories.")

ANOVA result: F-statistic = 155.8728154988859, p-value = 0.0
There is a significant difference in sales across Sub Categories.
In [56]:
# What is the sales distribution across different Profit/Loss categories?
sales_subcategory_fig = px.box(merged_data, x='Profit/Loss', y='Sales', color='Profit/Loss', height=600, width=800, title='Sales Distribution Across Profit/Loss Categories')
sales_subcategory_fig.show()
In [57]:
# T-test test to see if sales differs significantly across Profit/Loss categories
profit_loss_groups = [merged_data[merged_data['Profit/Loss'] == pl]['Sales'] for pl in merged_data['Profit/Loss'].unique()]
anova_profit_loss_result = stats.f_oneway(*profit_loss_groups)
print(f"ANOVA result: F-statistic = {anova_profit_loss_result.statistic}, p-value = {anova_profit_loss_result.pvalue}")
if anova_profit_loss_result.pvalue < 0.05:
    print("There is a significant difference in sales across Profit/Loss categories.")
else:
    print("No significant difference in sales across Profit/Loss categories.")

ANOVA result: F-statistic = 2.528365638963113, p-value = 0.11184761388544562
No significant difference in sales across Profit/Loss categories.
In [58]:
# What is the sales distribution across different age groups?
sales_age_fig = px.box(merged_data, x='Age Group', y='Sales', color='Age Group', height=600, width=800, title='Sales Distribution Across Age Groups')
sales_age_fig.show()
# Descriptive statistics for sales cross age groups
merged_data.groupby('Age Group', observed=True)['Sales'].describe()
Out[58]:
count mean std min 25% 50% 75% max
Age Group
18-30 2142.0 232.133459 693.009822 0.84 17.46 50.35 206.2825 17499.95
31-45 2837.0 221.491167 521.054030 0.85 17.71 56.70 203.9800 9099.93
46-60 2678.0 225.864731 484.825686 0.56 16.06 55.84 211.1225 5443.96
61+ 2337.0 242.505353 788.820189 0.44 17.54 54.66 217.0600 22638.48
In [59]:
# ANOVA test to see if sales differs significantly age group categories
age_group_sales = [merged_data[merged_data['Age Group'] == ag]['Sales'] for ag in merged_data['Age Group'].unique()]
anova_age_group_result = stats.f_oneway(*age_group_sales)
print(f"ANOVA result: F-statistic = {anova_age_group_result.statistic}, p-value = {anova_age_group_result.pvalue}")
if anova_age_group_result.pvalue < 0.05:
    print("There is a significant difference in sales across Age Group categories.")
else:
    print("No significant difference in sales across Age Group categories.")

ANOVA result: F-statistic = 0.5373061706953338, p-value = 0.6567015746561319
No significant difference in sales across Age Group categories.
In [60]:
# What is the sales distribution across different month groups?
# plotly.express.box does not accept a `size_max` parameter — remove it
sales_month_fig = px.box(merged_data, x='Month', y='Sales', color='Month', height=600, width=800, title='Sales Distribution Across Months')
sales_month_fig.show()
# Descriptive statistics for sales cross month groups
print(merged_data.groupby('Month')['Sales'].describe())
# ANOVA test to see if sales differs significantly across month categories
month_sales = [merged_data[merged_data['Month'] == month]['Sales'] for month in merged_data['Month'].unique()]
anova_month_result = stats.f_oneway(*month_sales)
print(f"ANOVA result: F-statistic = {anova_month_result.statistic}, p-value = {anova_month_result.pvalue}")
if anova_month_result.pvalue < 0.05:
    print("There is a significant difference in sales across Month categories.")
else:
    print("No significant difference in sales across Month categories.")

        count        mean          std   min      25%    50%       75%  \
Month                                                                    
1       381.0  249.146640   596.761641  1.73  16.4500  52.34  191.4700   
2       300.0  199.170867   574.458332  1.08  17.0625  60.43  203.9350   
3       696.0  294.548147  1143.775451  0.56  18.8175  59.15  213.2175   
4       668.0  206.230778   543.241925  1.19  15.5650  47.98  195.9325   
5       735.0  210.923578   523.488519  1.19  17.9400  49.57  189.0700   
6       717.0  212.996820   425.066112  0.44  17.3800  53.90  208.1600   
7       710.0  207.377620   458.628272  1.04  15.6275  58.40  234.1925   
8       706.0  225.274773   469.056880  1.36  16.7050  54.56  212.5175   
9      1383.0  222.451164   552.659766  0.88  15.5650  49.12  196.2300   
10      819.0  244.594664   866.241864  0.90  17.5400  51.71  202.6550   
11     1471.0  239.606451   573.098507  1.19  18.3550  59.94  230.3650   
12     1408.0  231.032344   521.839141  1.17  18.7150  56.30  220.0950   

            max  
Month            
1       5443.96  
2       8749.95  
3      22638.48  
4       9099.93  
5       8399.98  
6       4476.80  
7       8187.65  
8       4416.17  
9       9449.95  
10     17499.95  
11     10499.97  
12      9892.74  
ANOVA result: F-statistic = 1.157822390268044, p-value = 0.31108880559530716
No significant difference in sales across Month categories.
  • The sales distributions vary significantly across different categories. Technology has the highest median sales and the widest range, indicating higher sales values and greater variability. Furniture has moderate sales with a narrower range, while Office Supplies has the lowest median sales and the least variability. The ANOVA test confirms that these differences are statistically significant (p-value < 0.05).
  • The sub-categories with the highest average sales are Copiers ($2198.94), Machines ($1645.55), and Tables ($648.80). These categories likely represent higher-value products. In contrast, sub-categories like Fasteners ($13.94), Labels ($34.30), and Art ($34.07) have much lower average sales, indicating they are lower-value items. This information can help prioritize inventory and marketing efforts towards higher-performing sub-categories.
  • There are not significant differences in sales across customer segments, ship modes, regions, and profit/loss, and age group categories. In all these cases, the p-values from ANOVA tests are greater than 0.05, indicating that sales distributions are similar across these different groupings. This suggests that these attributes do not significantly impact sales in this dataset.
  • Overall, the analysis indicates that while customer age does not vary significantly across most attributes, sales performance is strongly influenced by product category and sub-category. This insight can guide strategic decisions in marketing, inventory management, and customer targeting to optimize sales outcomes.

Discount Analysis¶

In [61]:
# Discount distribution
plt.figure(figsize=(8, 5))
sns.histplot(merged_data['Discount'], bins=30, kde=True)
plt.title('Discount Distribution')
plt.xlabel('Discount')
plt.ylabel('Frequency')
plt.show()
No description has been provided for this image
In [62]:
# Box plot of discount
discount_box_fig = px.box(merged_data, y='Discount', height=600, width=800, title='Box Plot of Discount')
discount_box_fig.show()
In [63]:
# What is the discount distribution across different customer segments?
age_segment_fig = px.box(merged_data, x='Segment', y='Discount', color='Segment', height=600, width=800, title='Discount Distribution Across Customer Segments')
age_segment_fig.show()
In [64]:
# ANOVA test to see if discount differs significantly across customer segments
# Get the discount values for each segment
segment_discounts = [group['Discount'].values for name, group in merged_data.groupby('Segment')]

# Perform ANOVA
anova_result = stats.f_oneway(*segment_discounts)
anova_result
print(f"ANOVA result: F-statistic = {anova_result.statistic}, p-value = {anova_result.pvalue}")
if anova_result.pvalue < 0.05:
    print("There is a significant difference in discount across customer segments.")
else:
    print("No significant difference in discount across customer segments.")

ANOVA result: F-statistic = 2.0969225057441223, p-value = 0.12288792419848278
No significant difference in discount across customer segments.
In [65]:
# What is the discount distribution across different customer regions?
region_segment_fig = px.box(merged_data, x='Region', y='Discount', color='Region', height=600, width=800, title='Discount Distribution Across Customer Regions')
region_segment_fig.show()
# ANOVA test to see if discount differs significantly across customer regions
region_discounts = [group['Discount'].values for name, group in merged_data.groupby('Region')]
anova_region_discount_result = stats.f_oneway(*region_discounts)
print(f"ANOVA result: F-statistic = {anova_region_discount_result.statistic}, p-value = {anova_region_discount_result.pvalue}")
if anova_region_discount_result.pvalue < 0.05:
    print("There is a significant difference in discount across customer regions.")
else:
    print("No significant difference in discount across customer regions.")

ANOVA result: F-statistic = 2.3234618628111905, p-value = 0.07291493307877009
No significant difference in discount across customer regions.
In [66]:
# What is the discount distribution across different State categories?
state_segment_fig = px.box(merged_data, x='State', y='Discount', color='State', height=800, width=1000, title='Discount Distribution Across Different States')
state_segment_fig.show()
# Descriptive statistics for discount cross State categories
print(merged_data.groupby('State')['Discount'].describe())
# ANOVA test to see if discount differs significantly across states
state_discounts = [group['Discount'].values for name, group in merged_data.groupby('State')]
anova_state_discount_result = stats.f_oneway(*state_discounts)
print(f"ANOVA result: F-statistic = {anova_state_discount_result.statistic}, p-value = {anova_state_discount_result.pvalue}")
if anova_state_discount_result.pvalue < 0.05:
    print("There is a significant difference in discount across states.")
else:
    print("No significant difference in discount across states.")

                       count      mean       std  min  25%   50%     75%  max
State                                                                        
Alabama                127.0  0.178504  0.210031  0.0  0.0  0.20  0.2000  0.8
Arizona                301.0  0.184053  0.220745  0.0  0.0  0.20  0.2000  0.8
Arkansas                21.0  0.271429  0.266726  0.0  0.0  0.20  0.3000  0.7
California            1998.0  0.143093  0.196949  0.0  0.0  0.00  0.2000  0.8
Colorado               283.0  0.215618  0.231939  0.0  0.0  0.20  0.2000  0.8
Connecticut             74.0  0.194595  0.257433  0.0  0.0  0.20  0.2000  0.8
Delaware                93.0  0.141935  0.180773  0.0  0.0  0.10  0.2000  0.8
District of Columbia    11.0  0.100000  0.214476  0.0  0.0  0.00  0.1000  0.7
Florida                290.0  0.180414  0.225135  0.0  0.0  0.20  0.2000  0.8
Georgia                201.0  0.109950  0.190001  0.0  0.0  0.00  0.2000  0.8
Illinois               574.0  0.190575  0.228147  0.0  0.0  0.20  0.2000  0.8
Indiana                167.0  0.146707  0.209139  0.0  0.0  0.00  0.2000  0.8
Iowa                    45.0  0.104444  0.134765  0.0  0.0  0.00  0.2000  0.7
Kansas                  13.0  0.146154  0.210616  0.0  0.0  0.00  0.2000  0.7
Kentucky               149.0  0.112752  0.184277  0.0  0.0  0.00  0.2000  0.8
Louisiana               85.0  0.101176  0.153137  0.0  0.0  0.00  0.2000  0.7
Maryland                24.0  0.072917  0.140635  0.0  0.0  0.00  0.1625  0.6
Massachusetts          114.0  0.149123  0.206100  0.0  0.0  0.00  0.2000  0.8
Michigan               220.0  0.135545  0.189031  0.0  0.0  0.00  0.2000  0.8
Minnesota              167.0  0.121557  0.174546  0.0  0.0  0.00  0.2000  0.8
Mississippi             61.0  0.131148  0.175633  0.0  0.0  0.00  0.2000  0.8
Missouri                60.0  0.127500  0.208185  0.0  0.0  0.00  0.2000  0.8
Nebraska                32.0  0.134375  0.207301  0.0  0.0  0.00  0.2000  0.7
Nevada                  12.0  0.100000  0.104447  0.0  0.0  0.10  0.2000  0.2
New Hampshire           28.0  0.139286  0.189227  0.0  0.0  0.10  0.2000  0.7
New Jersey             123.0  0.153659  0.200529  0.0  0.0  0.20  0.2000  0.8
New Mexico              41.0  0.073171  0.146670  0.0  0.0  0.00  0.2000  0.8
New York              1110.0  0.123631  0.186555  0.0  0.0  0.00  0.2000  0.8
North Carolina         343.0  0.191983  0.225984  0.0  0.0  0.20  0.2000  0.8
Ohio                   480.0  0.176146  0.213356  0.0  0.0  0.20  0.2000  0.8
Oklahoma                16.0  0.037500  0.080623  0.0  0.0  0.00  0.0000  0.2
Oregon                  60.0  0.228333  0.265593  0.0  0.0  0.20  0.2250  0.8
Pennsylvania           641.0  0.183058  0.214056  0.0  0.0  0.20  0.2000  0.8
Rhode Island            64.0  0.126562  0.206438  0.0  0.0  0.00  0.2000  0.8
South Dakota            16.0  0.062500  0.158640  0.0  0.0  0.00  0.0000  0.6
Tennessee              167.0  0.156108  0.190989  0.0  0.0  0.20  0.2000  0.8
Texas                  917.0  0.182868  0.213297  0.0  0.0  0.20  0.2000  0.8
Utah                    80.0  0.166250  0.218710  0.0  0.0  0.05  0.2000  0.7
Virginia               182.0  0.147802  0.210945  0.0  0.0  0.00  0.2000  0.8
Washington             460.0  0.136717  0.191310  0.0  0.0  0.00  0.2000  0.8
Wisconsin              144.0  0.141319  0.199416  0.0  0.0  0.00  0.2000  0.8
ANOVA result: F-statistic = 5.25976445145791, p-value = 1.128827689411439e-24
There is a significant difference in discount across states.

State-wise discount analysis shows that states like Arkansas (mean discount: 27.14%), Colorado (21.56%), and Oregon (22.83%) offer higher average discounts compared to states like Maryland (7.29%) and Oklahoma (3.75%). The ANOVA test confirms that these differences are statistically significant (p-value < 0.05). This insight can help tailor discount strategies based on state-specific customer behavior.

  • Investigate whether higher discounts lead to increased sales and profitability in different states.
  • Top 5 states with highest average discounts: Arkansas, Oregon, Colorado, Arizona, North Carolina.
  • Bottom 5 states with lowest average discounts: Oklahoma, Maryland, South Dakota, Nevada, New Mexico.
In [67]:
# What is the discount distribution across different Ship Modes?
ship_mode_segment_fig = px.box(merged_data, x='Ship Mode', y='Discount', color='Ship Mode', height=600, width=800, title='Discount Distribution Across Different Ship Modes')
ship_mode_segment_fig.show()
# Descriptive statistics for discount cross Ship Mode categories
print(merged_data.groupby('Ship Mode')['Discount'].describe())
# ANOVA test to see if discount differs significantly across Ship Modes
ship_mode_discounts = [group['Discount'].values for name, group in merged_data.groupby('Ship Mode')]
anova_ship_mode_discount_result = stats.f_oneway(*ship_mode_discounts)
print(f"ANOVA result: F-statistic = {anova_ship_mode_discount_result.statistic}, p-value = {anova_ship_mode_discount_result.pvalue}")
if anova_ship_mode_discount_result.pvalue < 0.05:
    print("There is a significant difference in discount across Ship Modes.")
else:
    print("No significant difference in discount across Ship Modes.")

                 count      mean       std  min  25%  50%  75%  max
Ship Mode                                                          
First Class     1538.0  0.164610  0.210319  0.0  0.0  0.2  0.2  0.8
Same Day         543.0  0.152394  0.198365  0.0  0.0  0.2  0.2  0.8
Second Class    1945.0  0.138895  0.190728  0.0  0.0  0.0  0.2  0.8
Standard Class  5968.0  0.160023  0.210782  0.0  0.0  0.2  0.2  0.8
ANOVA result: F-statistic = 6.159416358787153, p-value = 0.0003531354423901233
There is a significant difference in discount across Ship Modes.
  • The discount distributions vary significantly across different ship modes. The ship modes with the highest average discounts are First Class (16.46%) and Standard Class (16.00%), while Second Class has the lowest average discount (13.89%). The ANOVA test confirms that these differences are statistically significant (p-value < 0.05). This insight can help optimize shipping strategies to balance cost and customer satisfaction.
  • The first class and standard class ship modes are higher discounts, possibly to attract customers and offset higher shipping costs.
In [68]:
# What is the discount distribution across different Categories?
category_segment_fig = px.box(merged_data, x='Category', y='Discount', color='Category', height=600, width=800, title='Discount Distribution Across Different Categories')
category_segment_fig.show()
# Descriptive statistics for discount cross Category categories
print(merged_data.groupby('Category')['Discount'].describe())
# ANOVA test to see if discount differs significantly across Categories
category_discounts = [group['Discount'].values for name, group in merged_data.groupby('Category')]
anova_category_discount_result = stats.f_oneway(*category_discounts)
print(f"ANOVA result: F-statistic = {anova_category_discount_result.statistic}, p-value = {anova_category_discount_result.pvalue}")
if anova_category_discount_result.pvalue < 0.05:
    print("There is a significant difference in discount across Categories.")
else:
    print("No significant difference in discount across Categories.")

                  count      mean       std  min  25%  50%  75%  max
Category                                                            
Furniture        2121.0  0.173923  0.181547  0.0  0.0  0.2  0.3  0.7
Office Supplies  6026.0  0.157285  0.229495  0.0  0.0  0.0  0.2  0.8
Technology       1847.0  0.132323  0.141575  0.0  0.0  0.2  0.2  0.7
ANOVA result: F-statistic = 20.329543937916956, p-value = 1.5449264084212215e-09
There is a significant difference in discount across Categories.

The category with the highest average discount is Furniture (17.39%), followed by Office Supplies (15.73%), and Technology (13.23%). The ANOVA test confirms that these differences are statistically significant (p-value < 0.05). This insight can help tailor discount strategies based on product categories to optimize sales and profitability.

  • Furniture items may have higher discounts to encourage purchases of typically lower-profit items. Or customers may expect higher discounts on furniture due to its higher price point and competitive market, cleaning out inventory.
  • Technology items have lower discounts, possibly due to higher base prices and profit margins.
In [69]:
# What is the discount distribution across different Sub Categories?
sub_category_segment_fig = px.box(merged_data, x='Sub-Category', y='Discount', color='Sub-Category', height=600, width=800, title='Discount Distribution Across Different Sub Categories')
sub_category_segment_fig.show()
# Descriptive statistics for discount cross Sub-Category categories
print(merged_data.groupby('Sub-Category')['Discount'].describe())
# ANOVA test to see if discount differs significantly across Sub-Category
sub_category_discounts = [group['Discount'].values for name, group in merged_data.groupby('Sub-Category')]
anova_sub_category_discount_result = stats.f_oneway(*sub_category_discounts)
print(f"ANOVA result: F-statistic = {anova_sub_category_discount_result.statistic}, p-value = {anova_sub_category_discount_result.pvalue}")
if anova_sub_category_discount_result.pvalue < 0.05:
    print("There is a significant difference in discount across Sub Categories.")
else:
    print("No significant difference in discount across Sub Categories.")

               count      mean       std  min   25%  50%   75%  max
Sub-Category                                                       
Accessories    775.0  0.078452  0.097714  0.0  0.00  0.0  0.20  0.2
Appliances     466.0  0.166524  0.272771  0.0  0.00  0.0  0.20  0.8
Art            796.0  0.074874  0.096853  0.0  0.00  0.0  0.20  0.2
Binders       1523.0  0.372292  0.310938  0.0  0.20  0.2  0.70  0.8
Books          228.0  0.211140  0.191913  0.0  0.00  0.2  0.32  0.7
Chairs         617.0  0.170178  0.107444  0.0  0.10  0.2  0.30  0.3
Copiers         68.0  0.161765  0.130475  0.0  0.00  0.2  0.20  0.4
Envelopes      254.0  0.080315  0.098237  0.0  0.00  0.0  0.20  0.2
Fasteners      217.0  0.082028  0.098599  0.0  0.00  0.0  0.20  0.2
Furnishings    957.0  0.138349  0.207804  0.0  0.00  0.0  0.20  0.6
Labels         364.0  0.068681  0.095100  0.0  0.00  0.0  0.20  0.2
Machines       115.0  0.306087  0.252804  0.0  0.05  0.2  0.50  0.7
Paper         1370.0  0.074891  0.096832  0.0  0.00  0.0  0.20  0.2
Phones         889.0  0.154556  0.129813  0.0  0.00  0.2  0.20  0.4
Storage        846.0  0.074704  0.096805  0.0  0.00  0.0  0.20  0.2
Supplies       190.0  0.076842  0.097539  0.0  0.00  0.0  0.20  0.2
Tables         319.0  0.261285  0.169622  0.0  0.20  0.3  0.40  0.5
ANOVA result: F-statistic = 216.20507931042397, p-value = 0.0
There is a significant difference in discount across Sub Categories.
ANOVA result: F-statistic = 216.20507931042397, p-value = 0.0
There is a significant difference in discount across Sub Categories.

The sub-categories with the highest average discounts are Machines (30.61%), Tables (26.13%), and Binders (37.23%). These categories likely represent higher-value products. In contrast, sub-categories like Labels (6.87%), Art (7.49%), and Storage (7.47%) have much lower average discounts, indicating they are lower-value items. This information can help prioritize inventory and marketing efforts towards higher-performing sub-categories.

  • Machines and Tables may have higher discounts to encourage purchases of typically higher-priced items.
  • Binders have high discounts, possibly to clear out inventory or attract bulk buyers.
  • Labels, Art, and Storage have lower discounts, possibly due to their lower price points and higher sales volumes.
In [70]:
# What is the discount distribution across different Profit/Loss categories?
profit_loss_segment_fig = px.box(merged_data, x='Profit/Loss', y='Discount', color='Profit/Loss', height=600, width=800, title='Discount Distribution Across Different Profit/Loss Categories')
profit_loss_segment_fig.show()
# Descriptive statistics for discount cross Profit/Loss categories
print(merged_data.groupby('Profit/Loss')['Discount'].describe())
# T-test test to see if discount differs significantly across Profit/Loss categories
profit_loss_discounts = [group['Discount'].values for name, group in merged_data.groupby('Profit/Loss')]
t_test_profit_loss_discount_result = stats.ttest_ind(*profit_loss_discounts)
print(f"T-test result: T-statistic = {t_test_profit_loss_discount_result.statistic}, p-value = {t_test_profit_loss_discount_result.pvalue}")
if t_test_profit_loss_discount_result.pvalue < 0.05:
    print("There is a significant difference in discount across Profit/Loss categories.")
else:
    print("No significant difference in discount across Profit/Loss categories.")

              count      mean       std  min  25%  50%  75%  max
Profit/Loss                                                     
Loss         1871.0  0.480887  0.235080  0.1  0.2  0.4  0.7  0.8
Profit       8123.0  0.081417  0.099192  0.0  0.0  0.0  0.2  0.4
T-test result: T-statistic = 115.02962858368716, p-value = 0.0
There is a significant difference in discount across Profit/Loss categories.

The loss category has a significantly higher average discount (48.09%) compared to the profit category (8.14%). The T-test confirms that this difference is statistically significant (p-value < 0.05). This insight suggests that higher discounts may be associated with unprofitable transactions, indicating a need to balance discount strategies with profitability goals.

  • The median discount for loss-making transactions is 40%, while for profitable transactions it is 0%, indicating that many profitable sales occur without discounts.
  • The maximum discount offered in loss-making transactions is 80%, which may be too high and negatively impact profitability.
  • Discount is the key factor differentiating profit and loss transactions in this dataset.
  • Reevaluating discount strategies could help improve overall profitability.
  • Focus on selling at low or no discount to increase profit margins, especially for high-demand products or high-value products.
In [71]:
# What is the discount distribution across different age groups?
discount_age_fig = px.box(merged_data, x='Age Group', y='Discount', color='Age Group', height=600, width=800, title='Discount Distribution Across Age Groups')
discount_age_fig.show()
# Descriptive statistics for discount across age groups
print(merged_data.groupby('Age Group', observed=True)['Discount'].describe())
# ANOVA test to see if discount differs significantly across age categories
age_discount = [merged_data[merged_data['Age Group'] == age]['Discount'] for age in merged_data['Age Group'].unique()]
anova_age_result = stats.f_oneway(*age_discount)
print(f"ANOVA result: F-statistic = {anova_age_result.statistic}, p-value = {anova_age_result.pvalue}")
if anova_age_result.pvalue < 0.05:
    print("There is a significant difference in discount across Age Group categories.")
else:
    print("No significant difference in discount across Age Group categories.")

            count      mean       std  min  25%  50%  75%  max
Age Group                                                     
18-30      2142.0  0.164220  0.210902  0.0  0.0  0.2  0.2  0.8
31-45      2837.0  0.158847  0.205954  0.0  0.0  0.2  0.2  0.8
46-60      2678.0  0.147483  0.203597  0.0  0.0  0.0  0.2  0.8
61+        2337.0  0.155635  0.205939  0.0  0.0  0.2  0.2  0.8
ANOVA result: F-statistic = 2.8318062735237133, p-value = 0.036862186381551275
There is a significant difference in discount across Age Group categories.

There is a significant difference in discount across Age Group categories.

  • For age groups 18-30 and 31-45, the average discounts are slightly higher (16.42% and 15.88% respectively) compared to older age groups 46-60 (14.75%) and 61+ (15.56%). This suggests that younger customers may be more responsive to discounts, potentially influencing their purchasing behavior.
In [72]:
# What is the discount distribution across different months?
month_segment_fig = px.box(merged_data, x='Month', y='Discount', color='Month', height=600, width=800, title='Discount Distribution Across Different Months')
month_segment_fig.show()
# ANOVA test to see if discount differs significantly across months
month_discounts = [group['Discount'].values for name, group in merged_data.groupby('Month')]
anova_month_discount_result = stats.f_oneway(*month_discounts)
print(f"ANOVA result: F-statistic = {anova_month_discount_result.statistic}, p-value = {anova_month_discount_result.pvalue}")
if anova_month_discount_result.pvalue < 0.05:
    print("There is a significant difference in discount across months.")
else:
    print("No significant difference in discount across months.")

ANOVA result: F-statistic = 0.7149143731674615, p-value = 0.7253874296623296
No significant difference in discount across months.

Overall, the discount analysis reveals significant variations in discount strategies across states, ship modes, product categories, age groups and profit/loss outcomes. These insights can inform targeted discounting approaches to enhance sales performance while maintaining profitability.

Profit Analysis¶

In [73]:
# What is the profit distribution across different customer segments?
profit_segment_fig = px.box(merged_data, x='Segment', y='Profit', color='Segment', height=600, width=800, title='Profit Distribution Across Customer Segments')
profit_segment_fig.show()
# Descriptive statistics for profit across customer segments
merged_data.groupby('Segment')['Profit'].describe()
Out[73]:
count mean std min 25% 50% 75% max
Segment
Consumer 5191.0 25.836897 242.594088 -6599.98 1.5850 8.38 28.610 6719.98
Corporate 3020.0 30.456772 231.986397 -3839.99 1.9275 8.69 29.845 8399.98
Home Office 1783.0 33.818850 212.381596 -3399.98 1.9800 9.10 31.560 3919.99
In [74]:
# ANOVA test to see if profit differs significantly across customer segments
anova_result = stats.f_oneway(
    merged_data[merged_data['Segment'] == 'Consumer']['Profit'],
    merged_data[merged_data['Segment'] == 'Corporate']['Profit'],
    merged_data[merged_data['Segment'] == 'Home Office']['Profit']
)
print(f"ANOVA result: F-statistic = {anova_result.statistic}, p-value = {anova_result.pvalue}")
if anova_result.pvalue < 0.05:
    print("There is a significant difference in profit across customer segments.")
else:
    print("No significant difference in profit across customer segments.")

ANOVA result: F-statistic = 0.8981014448131166, p-value = 0.4073751705325088
No significant difference in profit across customer segments.
In [75]:
# What is the profit distribution across different customer regions?
profit_region_fig = px.box(merged_data, x='Region', y='Profit', color='Region', height=600, width=800, title='Profit Distribution Across Customer Regions')
profit_region_fig.show()
# Descriptive statistics for profit across customer regions
merged_data.groupby('Region')['Profit'].describe()
Out[75]:
count mean std min 25% 50% 75% max
Region
Central 2371.0 26.828199 165.759190 -1862.31 1.5300 7.860 29.0550 5039.99
East 2762.0 34.252075 229.805819 -1665.05 1.7600 8.455 27.9900 6719.98
South 1626.0 18.558137 287.871799 -6599.98 1.7675 9.000 32.6125 3177.48
West 3235.0 30.296250 250.251927 -3399.98 1.9000 9.190 29.6350 8399.98
In [76]:
# ANOVA test to see if profit differs significantly across customer regions
anova_result = stats.f_oneway(
    merged_data[merged_data['Region'] == 'East']['Profit'],
    merged_data[merged_data['Region'] == 'West']['Profit'],
    merged_data[merged_data['Region'] == 'Central']['Profit']
)
print(f"ANOVA result: F-statistic = {anova_result.statistic}, p-value = {anova_result.pvalue}")
if anova_result.pvalue < 0.05:
    print("There is a significant difference in profit across customer regions.")
else:
    print("No significant difference in profit across customer regions.")

ANOVA result: F-statistic = 0.7168568381374917, p-value = 0.4883145981237146
No significant difference in profit across customer regions.
In [77]:
# What is the profit distribution across different customer regions?
profit_region_fig = px.box(merged_data, x='State', y='Profit', color='Region', color_discrete_sequence=px.colors.qualitative.Set1, height=800, width=1000, title='Profit Distribution Across Customer Regions')
profit_region_fig.show()
# Descriptive statistics for profit cross State categories
merged_data.groupby('State')['Profit'].describe()
Out[77]:
count mean std min 25% 50% 75% max
State
Alabama 127.0 22.402520 208.709685 -1237.85 0.0850 7.250 34.5350 1459.20
Arizona 301.0 31.771628 322.120448 -1181.28 0.0000 7.460 32.1000 4946.37
Arkansas 21.0 -2.998571 59.599418 -173.35 -7.4600 -1.220 11.8700 161.97
California 1998.0 29.728944 177.857678 -2929.48 2.5100 9.840 30.9875 4630.48
Colorado 283.0 3.429117 231.230585 -3399.98 -4.0350 6.510 24.1200 1270.99
Connecticut 74.0 7.209459 159.388797 -786.74 1.7850 10.270 29.2300 415.99
Delaware 93.0 35.875269 100.318964 -131.95 1.7600 9.870 46.1800 609.72
District of Columbia 11.0 44.633636 80.746584 -6.03 7.4400 9.330 22.1250 244.62
Florida 290.0 2.589276 229.769718 -3701.89 1.0375 7.510 24.0875 374.38
Georgia 201.0 63.589055 257.225107 -213.56 2.8300 11.690 40.4400 3177.48
Illinois 574.0 16.655418 113.742825 -938.28 0.5900 6.505 23.6975 1644.29
Indiana 167.0 16.212036 50.951695 -172.12 2.9450 9.480 26.4850 314.27
Iowa 45.0 50.191111 95.777671 -30.39 5.8900 14.240 51.7500 523.71
Kansas 13.0 10.708462 26.500089 -24.48 1.7100 11.630 15.4900 83.99
Kentucky 149.0 30.290671 80.398646 -216.98 2.7700 10.950 39.4300 353.58
Louisiana 85.0 31.285647 162.714028 -913.18 3.1100 8.640 30.2400 1007.98
Maryland 24.0 18.193750 24.504451 -4.80 4.6725 9.710 20.2850 103.23
Massachusetts 114.0 51.802982 260.035280 -302.72 2.5375 10.925 27.4200 2591.96
Michigan 220.0 35.237773 172.518894 -1862.31 2.7275 12.195 47.1850 829.38
Minnesota 167.0 43.129162 170.552080 -786.01 3.5500 9.640 32.5000 1013.13
Mississippi 61.0 9.027705 361.742704 -2639.99 2.1900 11.240 58.6900 607.61
Missouri 60.0 36.881167 125.245369 -320.24 3.9600 8.025 43.8775 757.41
Nebraska 32.0 36.438437 154.593887 -384.72 4.5200 9.750 32.7200 743.99
Nevada 12.0 23.172500 44.995018 1.67 4.5500 7.460 18.9925 163.82
New Hampshire 28.0 54.256071 120.497393 -70.49 3.7325 9.325 53.8025 485.99
New Jersey 123.0 18.998211 75.289949 -430.62 0.9950 8.000 27.4950 352.30
New Mexico 41.0 32.686585 44.536309 -9.56 5.2200 18.090 28.0100 173.24
New York 1110.0 52.412414 317.875647 -1306.55 2.8500 9.175 31.2375 6719.98
North Carolina 343.0 -12.351749 464.693226 -6599.98 1.1950 7.490 29.1500 1120.00
Ohio 480.0 12.470479 102.854824 -950.40 1.0300 6.655 21.7175 751.76
Oklahoma 16.0 51.814375 72.328343 0.70 4.7175 13.090 92.9525 236.23
Oregon 60.0 3.901167 294.828624 -1850.95 -4.3850 3.590 21.7475 1014.98
Pennsylvania 641.0 21.224509 156.807004 -1665.05 0.5900 6.720 25.1800 2302.97
Rhode Island 64.0 35.573437 91.223828 -87.15 2.4725 8.605 26.0400 481.87
South Dakota 16.0 42.660000 48.632135 -11.66 5.9875 23.825 65.2925 134.65
Tennessee 167.0 20.564850 128.511784 -1065.37 2.1550 9.070 32.2500 601.97
Texas 917.0 22.387099 202.903292 -1359.99 0.6000 6.370 22.6300 5039.99
Utah 80.0 22.727625 181.226375 -1002.78 0.3825 8.370 22.9300 1114.51
Virginia 182.0 38.132582 258.104585 -1480.03 2.0250 10.330 36.6650 2400.97
Washington 460.0 53.055957 429.922789 -653.28 2.1300 8.995 28.7325 8399.98
Wisconsin 144.0 59.513125 198.996169 -182.64 4.3850 12.370 40.3225 1453.12
In [78]:
# ANOVA test to see if profit differs significantly across customer regions
profit_groups = [merged_data[merged_data['Region'] == region]['Profit'].dropna() for region in merged_data['Region'].unique()]
anova_profit_result = stats.f_oneway(*profit_groups)
print(f"ANOVA result: F-statistic = {anova_profit_result.statistic}, p-value = {anova_profit_result.pvalue}")
if anova_profit_result.pvalue < 0.05:
    print("There is a significant difference in profit across customer regions.")
else:
    print("No significant difference in profit across customer regions.")

ANOVA result: F-statistic = 1.633742925742889, p-value = 0.17924706987207947
No significant difference in profit across customer regions.
In [79]:
# What is the profit distribution across different Ship Modes?
profit_ship_mode_fig = px.box(merged_data, x='Ship Mode', y='Profit', color='Ship Mode', height=600, width=800, title='Profit Distribution Across Ship Modes')
profit_ship_mode_fig.show()
# Descriptive statistics for profit across ship modes
merged_data.groupby('Ship Mode')['Profit'].describe()
Out[79]:
count mean std min 25% 50% 75% max
Ship Mode
First Class 1538.0 31.840020 257.794294 -2639.99 1.7925 8.78 28.8350 6719.98
Same Day 543.0 29.266851 226.689478 -3839.99 2.1550 8.03 29.5650 2229.02
Second Class 1945.0 29.535470 152.822282 -1862.31 2.6000 9.43 33.2200 2799.98
Standard Class 5968.0 27.494881 250.010674 -6599.98 1.4700 8.30 28.7625 8399.98
In [80]:
# ANOVA test to see if profit differs significantly across Ship Modes
anova_result = stats.f_oneway(
    merged_data[merged_data['Ship Mode'] == 'Standard Class']['Profit'],
    merged_data[merged_data['Ship Mode'] == 'Second Class']['Profit'],
    merged_data[merged_data['Ship Mode'] == 'First Class']['Profit'],
    merged_data[merged_data['Ship Mode'] == 'Same Day']['Profit']
)
print(f"ANOVA result: F-statistic = {anova_result.statistic}, p-value = {anova_result.pvalue}")
if anova_result.pvalue < 0.05:
    print("There is a significant difference in profit across Ship Modes.")
else:
    print("No significant difference in profit across Ship Modes.")

ANOVA result: F-statistic = 0.1539104331482834, p-value = 0.9272120707211975
No significant difference in profit across Ship Modes.
In [81]:
# What is the profit distribution across different Categories?
profit_category_fig = px.box(merged_data, x='Category', y='Profit', color='Category', height=600, width=800, title='Profit Distribution Across Categories')
profit_category_fig.show()
# Descriptive statistics for profit across categories
merged_data.groupby('Category')['Profit'].describe()
Out[81]:
count mean std min 25% 50% 75% max
Category
Furniture 2121.0 8.699316 136.049323 -1862.31 -12.8500 7.770 33.7300 1013.13
Office Supplies 6026.0 20.327063 164.887346 -3701.89 2.1025 6.885 19.4175 4946.37
Technology 1847.0 78.752388 428.816759 -6599.98 5.2000 25.020 74.8950 8399.98
In [82]:
# ANOVA test to see if profit differs significantly across categories
anova_result = stats.f_oneway(
    merged_data[merged_data['Category'] == 'Furniture']['Profit'],
    merged_data[merged_data['Category'] == 'Office Supplies']['Profit'],
    merged_data[merged_data['Category'] == 'Technology']['Profit']
)
print(f"ANOVA result: F-statistic = {anova_result.statistic}, p-value = {anova_result.pvalue}")
if anova_result.pvalue < 0.05:
    print("There is a significant difference in profit across categories.")
else:
    print("No significant difference in profit across categories.")

ANOVA result: F-statistic = 54.311638544183815, p-value = 3.467806223383911e-24
There is a significant difference in profit across categories.
In [83]:
# What is the profit distribution across different Sub Categories?
profit_sub_category_fig = px.box(merged_data, x='Sub-Category', y='Profit', color='Sub-Category', height=600, width=800, title='Profit Distribution Across Sub Categories')
profit_sub_category_fig.show()
# Descriptive statistics for profit across sub categories
merged_data.groupby('Sub-Category')['Profit'].describe()
Out[83]:
count mean std min 25% 50% 75% max
Sub-Category
Accessories 775.0 54.111974 106.155481 -75.60 5.7750 21.000 57.3550 829.38
Appliances 466.0 38.922897 148.319322 -1181.28 4.0650 17.450 56.1125 793.72
Art 796.0 8.200955 13.384232 0.15 1.6850 3.720 8.6750 112.57
Binders 1523.0 19.843493 306.196875 -3701.89 -6.0250 3.980 15.8350 4946.37
Books 228.0 -15.230526 182.034336 -1665.05 -42.6325 4.135 40.0550 1013.13
Chairs 617.0 43.095867 127.099999 -630.88 -14.4800 13.320 70.2000 770.35
Copiers 68.0 817.910294 1460.921556 60.00 164.2475 332.990 848.7400 8399.98
Envelopes 254.0 27.417717 35.354703 0.55 5.9050 12.715 32.5375 204.07
Fasteners 217.0 4.375714 5.054964 -11.83 0.8700 2.840 6.2300 21.89
Furnishings 957.0 13.646029 52.753256 -427.45 2.7900 9.100 21.8200 387.57
Labels 364.0 15.236758 34.084317 0.68 3.4650 6.870 13.3025 385.38
Machines 115.0 29.432435 1099.070271 -6599.98 -104.6100 39.000 219.5250 2799.98
Paper 1370.0 24.856453 35.225415 1.07 6.1625 11.540 26.8800 352.30
Phones 889.0 50.074522 113.251222 -386.39 4.9600 23.520 67.5000 1228.18
Storage 846.0 25.152541 81.923647 -337.81 0.8500 7.755 24.8550 792.27
Supplies 190.0 -6.257842 105.059509 -1049.34 0.4400 3.960 9.0825 327.51
Tables 319.0 -55.566113 230.565048 -1862.31 -122.8300 -31.370 33.8500 629.01
In [84]:
# ANOVA test to see if profit differs significantly across sub categories
profit_sub_category_groups = [merged_data[merged_data['Sub-Category'] == subcategory]['Profit'] for subcategory in merged_data['Sub-Category'].unique()]
anova_profit_sub_category_result = stats.f_oneway(*profit_sub_category_groups)
print(f"ANOVA result: F-statistic = {anova_profit_sub_category_result.statistic}, p-value = {anova_profit_sub_category_result.pvalue}")
if anova_profit_sub_category_result.pvalue < 0.05:
    print("There is a significant difference in profit across Sub Categories.")
else:
    print("No significant difference in profit across Sub Categories.")

ANOVA result: F-statistic = 58.82082180393576, p-value = 2.81154065598055e-181
There is a significant difference in profit across Sub Categories.
In [85]:
# What is the profit distribution across different age groups?
profit_age_group_fig = px.box(merged_data, x='Age Group', y='Profit', color='Age Group', height=600, width=800, title='Profit Distribution Across Age Groups')
profit_age_group_fig.show()
# Descriptive statistics for profit across age groups
merged_data.groupby('Age Group', observed=True)['Profit'].describe()
Out[85]:
count mean std min 25% 50% 75% max
Age Group
18-30 2142.0 31.706261 306.228841 -3839.99 1.44 7.990 27.290 8399.98
31-45 2837.0 20.505048 204.789703 -6599.98 1.68 8.950 28.760 2365.98
46-60 2678.0 30.119679 163.426250 -1862.31 2.15 9.275 32.265 2504.22
61+ 2337.0 34.082011 259.164638 -1850.95 1.68 8.470 30.100 6719.98
In [86]:
# ANOVA test to see if profit differs significantly across age group categories
profit_age_group_groups = [merged_data[merged_data['Age Group'] == age_group]['Profit'] for age_group in merged_data['Age Group'].unique()]
anova_profit_age_group_result = stats.f_oneway(*profit_age_group_groups)
print(f"ANOVA result: F-statistic = {anova_profit_age_group_result.statistic}, p-value = {anova_profit_age_group_result.pvalue}")
if anova_profit_age_group_result.pvalue < 0.05:
    print("There is a significant difference in profit across Age Groups.")
else:
    print("No significant difference in profit across Age Groups.")

ANOVA result: F-statistic = 1.7190748073093414, p-value = 0.16071359445775768
No significant difference in profit across Age Groups.
In [87]:
# What is the profit distribution across different month groups?
profit_month_fig = px.box(merged_data, x='Month', y='Profit', color='Month', height=600, width=800, title='Profit Distribution Across Months')
profit_month_fig.show()
# Descriptive statistics for profit across month groups
print(merged_data.groupby('Month')['Profit'].describe())
# ANOVA test to see if profit differs significantly across month categories
month_profit = [merged_data[merged_data['Month'] == month]['Profit'] for month in merged_data['Month'].unique()]
anova_month_profit_result = stats.f_oneway(*month_profit)
print(f"ANOVA result: F-statistic = {anova_month_profit_result.statistic}, p-value = {anova_month_profit_result.pvalue}")
if anova_month_profit_result.pvalue < 0.05:
    print("There is a significant difference in profit across Month categories.")
else:
    print("No significant difference in profit across Month categories.")

        count       mean         std      min     25%     50%      75%  \
Month                                                                    
1       381.0  23.975013  221.694901 -1862.31  1.4800   7.600  26.4000   
2       300.0  34.315400  190.721837  -426.99  1.8825   8.475  30.6825   
3       696.0  41.084253  329.746292 -1811.08  1.5875   8.685  31.9350   
4       668.0  17.346662  219.361917 -3399.98  1.3625   6.555  24.2475   
5       735.0  30.491646  123.996040  -458.15  1.7750   9.100  27.0350   
6       717.0  29.687406  122.952811  -525.64  1.6800   9.020  31.0100   
7       710.0  19.482746  173.705345 -3701.89  1.8475   8.315  33.6650   
8       706.0  30.845708  135.147295 -1237.85  2.4000  10.585  33.5175   
9      1383.0  26.650398  192.928592 -1665.05  2.0550   8.280  28.5500   
10      819.0  38.808547  370.590587 -1306.55  1.1550   7.480  23.0450   
11     1471.0  24.111829  286.967683 -6599.98  1.4900   8.990  29.4050   
12     1408.0  30.802010  222.342045 -2929.48  2.1300   9.330  33.8775   

           max  
Month           
1      2504.22  
2      2799.98  
3      6719.98  
4      2365.98  
5      1276.49  
6      1668.21  
7       944.99  
8      1453.12  
9      4630.48  
10     8399.98  
11     5039.99  
12     4946.37  
ANOVA result: F-statistic = 0.6691946575058476, p-value = 0.7690778496956812
No significant difference in profit across Month categories.
  • The profit distributions vary significantly across different categories. Technology has the highest median profit and the widest range, indicating higher profit values and greater variability. Furniture has moderate profit with a narrower range, while Office Supplies has the lowest median profit and the least variability. The ANOVA test confirms that these differences are statistically significant (p-value < 0.05).
  • The sub-categories with the highest average profit are Copiers ($817.91), Accessories ($54.11), and Phones ($50.07). These categories likely represent higher-margin products. In contrast, sub-categories like Tables (-$55.57), Books (-$15.23), and Supplies (-$6.26) have negative average profits, indicating they are loss-making items. This information can help prioritize inventory and marketing efforts towards higher-performing sub-categories. Another observation is that there are not significant differences in profit across customer segments, ship modes, regions, and age group categories. In all these cases, the p-values from ANOVA tests are greater than 0.05, indicating that profit distributions are similar across these different groupings. This suggests that these attributes do not significantly impact profit in this dataset.
  • Overall, the analysis indicates that while customer age does not vary significantly across most attributes, sales and profit performance is strongly influenced by product category and sub-category. This insight can guide strategic decisions in marketing, inventory management, and customer targeting to optimize profit outcomes.
In [88]:
### Sales vs Profit with categories
high_selling_sub_categories = merged_data.groupby('Category')[['Sales', 'Profit']].sum()
high_selling_sub_categories = high_selling_sub_categories.reset_index()
top_sub_melted = high_selling_sub_categories.melt(id_vars='Category', value_vars=['Sales', 'Profit'], var_name='Metric', value_name='Value')
high_selling_sub_categories_fig = px.bar(top_sub_melted, x='Category', y='Value', color='Metric',height=600, width=800, text_auto=True, title='Total Sales and Profit by Category', labels={'x': 'Category', 'y': 'Total Sales'})
high_selling_sub_categories_fig.show()

Sales vs Profit with sub-categories¶

In [89]:
# High-selling sub-categories
high_selling_sub_categories = merged_data.groupby('Sub-Category')[['Sales', 'Profit']].sum()
high_selling_sub_categories = high_selling_sub_categories.reset_index()
top_sub_melted = high_selling_sub_categories.melt(id_vars='Sub-Category', value_vars=['Sales', 'Profit'], var_name='Metric', value_name='Value')
high_selling_sub_categories_fig = px.bar(top_sub_melted, x='Sub-Category', y='Value', color='Metric',height=1000, width=1400, text_auto=True, title='Total Sales and Profit by Sub-Category', labels={'x': 'Sub-Category', 'y': 'Total Sales'})
high_selling_sub_categories_fig.show()
In [90]:
# Scatter plot total profit vs quantity with sales size by sub-category
profit_sales_by_subcategory = merged_data.groupby('Sub-Category', observed=True).agg({'Profit': 'sum', 'Quantity': 'sum', 'Sales': 'sum'}).reset_index()
profit_sales_by_subcategory['Label'] = profit_sales_by_subcategory['Sub-Category'] + '<br>Quantity: ' + profit_sales_by_subcategory['Quantity'].astype(str)
# Create scatter plot
fig = px.scatter(
    profit_sales_by_subcategory,
    x='Sales',
    y='Profit',
    color='Sub-Category',
    color_discrete_sequence=px.colors.qualitative.Vivid, 
    height=1000, 
    width=1400,
    size='Quantity',
    size_max=80,
    text='Label',
    title='Total Profit vs Quantity by Sub-Category',
    labels={'Quantity': 'Total Quantity', 'Profit': 'Total Profit'}
)
fig.update_layout(yaxis_tickformat='$,.0f',
                  xaxis_tickformat='$,.0f')
fig.add_hline(y=profit_sales_by_subcategory['Profit'].mean(), line_dash="dash", line_color="gray",
              annotation_text="Average Profit", annotation_position="top left")
fig.add_hline(y=0, line_dash="dot", line_color="red",
              annotation_text="Break-even Line", annotation_position="bottom left")
fig.add_vline(x=profit_sales_by_subcategory['Sales'].mean(), line_dash="dash", line_color="gray",
              annotation_text="Average Sales", annotation_position="top right")
fig.show()

Key Observations¶

  • Phone high quantity and very high profit, this sub-category is a clear profit driver.
  • Chairs have a high quantity, but profit is much lower compared to phones, an opportunity to improve profit margin on chairs.
  • Tables quantity but negative profit, need to investigate reasons for losses and consider strategies to improve profitability or discontinue. This is a red flag for the business, even though sales are high. They are losing money on this sub-category.
  • Binders and Paper: high quantities sold but relatively modest profit, opportunity to optimize pricing or cost structure to boost profitability.
  • Copiers have low quantity but very high profit, focus on marketing and sales efforts to maximize revenue from this sub-category. This is a premium product with a high profit margin.
  • Accessories, Storage, and Furnishings are mid-tier items with moderate quantities and profits; these are steady contributors to overall profitability.
  • Fasteners, Envelopes, Art, and Machines, with low quantities and low profits, may not warrant significant focus but should be reviewed periodically to ensure they do not negatively impact overall profitability.
  • Books' negative profit should be carefully analyzed to identify reasons for losses and consider strategies such as price adjustments, cost reductions, or discontinuation if profitability cannot be improved.

Key Recommendations¶

  • Focus on Phones and Copiers as key profit drivers.
  • Investigate and address losses in Tables, Supplies, and Books sub-categories.
  • Optimize Chairs has good sales volume but a weaker profit margin. Explore pricing or cost strategies to boost profitability.
  • Leverage Binders and Paper for steady revenue, while monitoring mid-tier items for sales trends.
  • Reassess low performers like Fasteners, Envelopes, Art, and Machines may not justify significant focus, heavy marketing, or inventory efforts.
In [91]:
# Scatter plot average profit vs average discount with sales size by sub-category
avg_profit_discount_by_subcategory = merged_data.groupby('Sub-Category', observed=True).agg({'Profit': 'sum', 'Discount': 'mean', 'Quantity': 'sum'}).reset_index()
print(avg_profit_discount_by_subcategory)
avg_profit_discount_by_subcategory['Label'] = avg_profit_discount_by_subcategory['Sub-Category'] + '<br>Quantity: ' + avg_profit_discount_by_subcategory['Quantity'].astype(str)
# Create scatter plot
fig = px.scatter(
    avg_profit_discount_by_subcategory,
    x='Discount',
    y='Profit',
    color='Sub-Category',
    color_discrete_sequence=px.colors.qualitative.Vivid, 
    height=1000, 
    width=1400,
    size='Quantity',
    size_max=80,
    text='Label',
    title='Average Profit vs Average Discount by Sub-Category',
    labels={'Quantity': 'Total Quantity', 'Profit': 'Average Profit'}
)
fig.update_layout(yaxis_tickformat='$,.0f',
                  xaxis_tickformat='.0%')
fig.add_hline(y=avg_profit_discount_by_subcategory['Profit'].mean(), line_dash="dash", line_color="gray",
              annotation_text="Average Profit", annotation_position="top left")
fig.add_hline(y=0, line_dash="dot", line_color="red",
              annotation_text="Break-even Line", annotation_position="bottom left")
fig.add_vline(x=avg_profit_discount_by_subcategory['Discount'].mean(), line_dash="dash", line_color="gray",
              annotation_text="Average Discount", annotation_position="top right")
fig.show()

   Sub-Category    Profit  Discount  Quantity
0   Accessories  41936.78  0.078452      2976
1    Appliances  18138.07  0.166524      1729
2           Art   6527.96  0.074874      3000
3       Binders  30221.64  0.372292      5974
4         Books  -3472.56  0.211140       868
5        Chairs  26590.15  0.170178      2356
6       Copiers  55617.90  0.161765       234
7     Envelopes   6964.10  0.080315       906
8     Fasteners    949.53  0.082028       914
9   Furnishings  13059.25  0.138349      3563
10       Labels   5546.18  0.068681      1400
11     Machines   3384.73  0.306087       440
12        Paper  34053.34  0.074891      5178
13       Phones  44516.25  0.154556      3289
14      Storage  21279.05  0.074704      3158
15     Supplies  -1188.99  0.076842       647
16       Tables -17725.59  0.261285      1241

Key Observations¶

  • Copiers' profit is very high (55617.90), low quantity sold (234), discount at 16% is a high margin product, but low sales volume.
  • Phone's profit is high (44516.25), high quantity sold (3289), discount at 15% good balance of volume and margin. Strong balance of sales and profit.
  • Accessories profit is good (41936.78), high quantity sold (2976), discount at 8% solid performer. A healthy margin product with good sales volume.
  • Paper profit is good (34053.34), very high quantity sold (5178), discount at 7.5% high volume, low margin product. A volume driver with solid profit contribution.

=> These are profit drivers for the business.

  • Binders' profit is good (30221.64), very high quantity sold (5974), discount at 37% very high discount to drive volume, but still profitable. The high discount strategy is working here. It is heavily discount-driven. Without a discount, it may be loss-making.
  • Chair's profit is moderate (26590.15), high quantity sold (2356), discount at 17% opportunity to improve margin on good sales volume.

=> These are volume drivers but discount-sensitive. Test smaller discounts to see demand impact.

  • Storage profit is moderate (21279.05), high quantity sold (3158), discount at 7.5% solid performer.
  • Furnishings profit is moderate (13059.25), moderate quantity sold (3563), discount at 14% steady contributor.
  • Appliances profit is moderate (18138.07), moderate quantity sold (1729), discount at 16% steady contributor.
  • Art profit is low (6527.96), moderate quantity sold (3000), discount at 7.5% low performer.

=> These categories sell decently but are not major profit contributors. Monitor for trends. They may need to optimize costs or bundle with other products to boost sales.

  • The table's profit is negative (-17725.59), a moderate quantity was sold (1241), and a discount of 26% needs to be investigated to determine the reasons for the losses and consider strategies to improve profitability or discontinue.
  • Book's profit is negative (-3472.56), low quantity sold (868), discount at 21% need to investigate reasons for losses and consider strategies to improve profitability or discontinue.
  • Supplies profit is negative (-1188.99), low quantity sold (647), discount at 7.5% need to investigate reasons for losses and consider strategies to improve profitability or discontinue.

=> These are loss-making sub-categories that require immediate attention. Analyze cost structure, pricing, and demand to identify improvement strategies or consider discontinuation. The problems here could be high costs, low prices, or insufficient demand.

  • Feasters, Labels, Evenlopes, and Machines have low profit but are not loss-making, with moderate to low quantity sold. These may not warrant significant focus, but should be reviewed periodically to ensure they do not negatively impact overall profitability.

Key Recommendations¶

  • Protect Stars: Focus on maximizing sales and marketing efforts for Copiers, Phones, Accessories, and Paper to leverage their substantial profit and sales performance. Keep low discounts, invest in supply chain efficiency to maintain healthy margins.
  • Optimize Cash Cows: Review discount strategies for Binders and Chairs to improve profit margins without significantly impacting sales volume. Experiment with lower discounts to find the optimal balance.
  • Monitor Steady Contributors: Keep an eye on Storage, Furnishings, Appliances, and Art for any changes in sales trends. Consider bundling or promotional strategies to boost their performance.
  • Address Loss-Makers: Conduct a thorough analysis of Tables, Books, and Supplies to identify the root causes of losses. Explore options such as price adjustments, cost reductions, or discontinuation if profitability cannot be improved.
  • Review Low Performers: Periodically assess Fasteners, Labels, Envelopes, and Machines to ensure they do not negatively impact overall profitability. Consider phasing out if they consistently underperform.
In [92]:
# Products with highest profit
top_profit_products = merged_data.groupby('Product ID', observed=True)['Profit'].sum().sort_values(ascending=False)
print("Top 10 products with highest profit:")
print(top_profit_products.head(10))
print("Bottom 10 products with lowest profit:")
print(top_profit_products.tail(10))

Top 10 products with highest profit:
Product ID
TEC-CO-10004722    25199.94
OFF-BI-10003527     7753.06
TEC-CO-10001449     6983.89
TEC-CO-10003763     4570.94
TEC-AC-10002049     4425.32
TEC-MA-10001127     4094.98
TEC-MA-10003979     3772.95
TEC-MA-10001047     3717.97
OFF-BI-10001120     3345.29
TEC-MA-10000045     3343.53
Name: Profit, dtype: float64
Bottom 10 products with lowest profit:
Product ID
FUR-TA-10004289   -1148.44
FUR-TA-10001950   -1201.06
OFF-SU-10002881   -1299.19
TEC-MA-10002412   -1811.08
OFF-BI-10004995   -1878.17
FUR-TA-10001889   -1934.40
FUR-TA-10000198   -2876.11
TEC-MA-10004125   -3839.99
TEC-MA-10000822   -4589.97
TEC-MA-10000418   -8879.97
Name: Profit, dtype: float64

In summary, the profit analysis highlights the need to focus on high-performing sub-categories while addressing the challenges faced by loss-making items. Strategic adjustments in pricing, discounting, and marketing can help optimize overall profitability.

Discount Analysis¶

  • The discount analysis reveals significant variations in discount strategies across states, ship modes, product categories, age groups and profit/loss outcomes. These insights can inform targeted discounting approaches to enhance sales performance while maintaining profitability. The category and sub-category analyses provide actionable recommendations for optimizing inventory management, marketing strategies, and overall business performance.

Key Findings¶

  • States like Arkansas, Oregon, and Colorado offer higher average discounts, while states like Maryland and Oklahoma have lower discounts.
  • First Class and Standard Class ship modes have higher average discounts compared to Second Class.
  • Furniture category has the highest average discount, while Technology has the lowest.
  • Loss-making transactions have significantly higher average discounts compared to profitable ones.
  • Younger age groups (18-30 and 31-45) tend to receive slightly higher discounts compared to older age groups (46-60 and 61+).

Summary of Findings¶

Overall, the discount analysis reveals significant variations in discount strategies across states, ship modes, product categories, age groups and profit/loss outcomes. These insights can inform targeted discounting approaches to enhance sales performance while maintaining profitability. The category and sub-category analyses provide actionable recommendations for optimizing inventory management, marketing strategies, and overall business performance. Profitability appears to be strongly influenced by product category and sub-category, while customer demographics such as age, segment, ship mode, and region have less impact.

Suggested Next Steps¶

  • Further investigate the impact of discounts on sales volume and profitability to refine discount strategies.
  • Explore customer feedback and preferences regarding discounts to better tailor offerings.
  • Conduct A/B testing of different discount strategies to identify the most effective approaches.
  • The discount distributions vary significantly across different ship modes. The ship modes with the highest average discounts are First Class (16.46%) and Standard Class (16.00%), while Second Class has the lowest average discount (13.89%). The ANOVA test confirms that these differences are statistically significant (p-value < 0.05). This insight can help optimize shipping strategies to balance cost and customer satisfaction.
  • The first class and standard class ship modes are higher discounts, possibly to attract customers and offset higher shipping costs.
  • Negative profit items like Tables, Supplies, and Book should be carefully analyzed to identify reasons for losses. Consider strategies such as price adjustments, cost reductions, or discontinuation if profitability cannot be improved. Review pricing strategies, source alternatives, and marketing approaches to mitigate losses.
  • High profit-low quantity items like Copiers should be prioritized for marketing and sales efforts to maximize revenue.
  • High profit-high quantity items like Paper, Accessories, Phones and Binders should be stocked adequately to meet demand and capitalize on their profitability. Prioritize in promotions, discounts, and advertising campaigns. Ensure sufficient inventory levels to avoid stockouts, and explore cross-selling opportunities with related products.
  • Mid-Tier items like Chairs, Furnishings, Appliances, and Storage should be monitored for sales trends and adjusted inventory levels accordingly. Consider bundling with high-profit items to boost sales.